Papers by Tomoki Hamagami
Keisoku Jidō Seigyo Gakkai ronbunshū, 2008
This paper proposes a technique for realizing cooperation behavior emerging among intelligent whe... more This paper proposes a technique for realizing cooperation behavior emerging among intelligent wheel chairs (IWC). In order to acquire the behavior in various environments, a system with a broadcasting framework and a decision tree is implemented, which uses a frustration level indicating a dynamical internal state of an IWC. The system brings diversity among each IWC's, and resolves a deadlock of them. Moreover, in order to design an adaptive behavior, the decision tree is generated with the genetic programming. As results of experiment with simulation and real prototype IWC, we confirm the mutual concession behavior emerges in narrow aisle and cross road environment.
The transactions of the Institute of Electrical Engineers of Japan.C, 1999
This paper proposes a new applied method to compress the VCV (Vowel-Consonant-Vowel) speech spec ... more This paper proposes a new applied method to compress the VCV (Vowel-Consonant-Vowel) speech spec trum pattern for synthesis by rule using a Genetic Algorithm (GA). In this method synthesized speech is
IEICE Transactions on Information and Systems, Mar 1, 2023
Ordinal regression is used to classify instances by considering ordinal relation between labels. ... more Ordinal regression is used to classify instances by considering ordinal relation between labels. Existing methods tend to decrease the accuracy when they adhere to the preservation of the ordinal relation. Therefore, we propose a distributional knowledge-based network (DK-net) that considers ordinal relation while maintaining high accuracy. DK-net focuses on image datasets. However, in industrial applications, one can find not only image data but also tabular data. In this study, we propose DK-neural oblivious decision ensemble (NODE), an improved version of DK-net for tabular data. DK-NODE uses NODE for feature extraction. In addition, we propose a method for adjusting the parameter that controls the degree of compliance with the ordinal relation. We experimented with three datasets: WineQuality, Abalone, and Eucalyptus dataset. The experiments showed that the proposed method achieved high accuracy and small MAE on three datasets. Notably, the proposed method had the smallest average MAE on all datasets.
Springer eBooks, 2012
Recently, various applications and services start to be used in the Internet. Load balancing the ... more Recently, various applications and services start to be used in the Internet. Load balancing the increasing network traffic in real time can affect the network quality. The flow control technologies become much more important than before. Our research project proposes an intelligent network flow identifying method, smart flow, which is based on the learning algorithm. In this paper, we suggest to utilize the SOM for learning the properties of packets, such as timestamp, source and destination. Based on our proposed normalization, IP network flows can be formed autonomously during the learning process. Furthermore, the combination use of the new normalization with the GHSOM can classify the sub-IP flows belongs to the same flow. This paper indicates that a flow shall consist of several sub-IP flows, and sub-IP flow shall consist of several IP packets.
Keisoku Jidō Seigyo Gakkai ronbunshū, 2009
This paper proposes two kinds of complex-valued Profit Sharing algorithm for learning in environm... more This paper proposes two kinds of complex-valued Profit Sharing algorithm for learning in environments involving perceptual aliasings. These algorithms are implementations in complex-valued reinforcement learning framework. The CVRL framework aims to solve perceptual aliasing problem by using context. The CVRL is characterized by complex-valued action values and a internal reference value. The internal reference value is used to distinguish confused states and revised at each step. As another implementation in the framework, Q-learning has been proposed. Several experimental results have demonstrated the effectiveness of the Q-learning. However, Q-learning suffers from a disadvantage in parameter settings because of many parameters. Therefore, this paper focuses on Profit Sharing which has less parameter than Q-learning and proposes complex-valued Profit Sharing. Furthermore, this paper also describes a technique named multiple phase changes. This technique enables the agent to revise the internal reference value adaptively and realizes more sufficient learning in environments which have confused states at various intervals. Simulation results support that proposed method works well in such environments.
Ieej Transactions on Electrical and Electronic Engineering, 2008
The goal of this study is to develop heterogeneous mobile robots, which realize cooperative actio... more The goal of this study is to develop heterogeneous mobile robots, which realize cooperative actions to accomplish each task in an environment. An important part of achieving the goal is that the agent controlling the robots can realize suitable state space automatically. In particular, domestic robots equipped with low-cost, discrete simple range sensors have difficulties in the efficient learning of environment because the deviation of sensor configuration in each robot causes serious problems in which the learning policies cannot be shared among them. To overcome this problem, a new method of constructing effective state space with low dependence on sensor configurations is conducted. The method consists of two parts: (i) abstracting surrounding environment by rectangles, and (ii) generalization of state space by classifying with self organization map (SOM). Simulation experiments show that the state space that has been constructed by an agent with 10 sensors can be reused for the agent with another sensor configuration.
The transactions of the Institute of Electrical Engineers of Japan.C, Sep 1, 2019
High performance Block-Based Neural Network Model by pipelined parallel communication Kundo Lee *... more High performance Block-Based Neural Network Model by pipelined parallel communication Kundo Lee * , * * a) , Non-member, Tomoki Hamagami * * , Member The structure and weight in Block-Based Neural Network (BBNN) are optimized by utilizing genetic algorithm. The architecture of BBBN consists of a two-dimensional (2-D) array of basic block with four input/output nodes and connection weights for block's output. To propose easier hardware implementation like Field Programmable Gate Array (FPGA), integer weights are used in the basic block. Each block can be one of the four different basic types and the architecture of BBNN is configured with the combination of basic block internally configured. However, BBNN's structural change needs hardware reconfiguration and the cost is very high. To reduce the reconfiguration cost, Smart Block-based Neuron (SBbN) which has sufficient number of weights for all four types of basic block has been proposed. SBbN preserves all weights even unnecessary for some types, and thus it consumes redundant hardware resource. A new model of BBNNs in which all weights in SBbN are used efficiently with modifying calculation procedures of outputs of basic blocks has been proposed and it eliminates the resource redundancy of SBbN. However, new approach which both, left and right's side nodes concurrently serve as input and output does not provide parallel computation in left and rightward signal flow. This paper presents a parallel computation with independent side nodes for each signal flow.
The transactions of the Institute of Electrical Engineers of Japan.C, 2012
A new call-triage system, a key part of emergency support system with stochastic network model is... more A new call-triage system, a key part of emergency support system with stochastic network model is examined. The call-triage is an operation allowing the efficient decision of service grade and dispatching of suitable rescue team service from phone call information. Nowadays, the call-triage is being trialed on a few cities and is achieving an effect. However, there is the issue that if under-triage in which the condition of sick person is estimated more lightly is eliminated, the efficiency is degraded (over-triage). In this report, in order to overcome the issue, the Bayesian network scheme is examined to the call-triage system. The experiments with real call-triage data set results show the Bayesian network achieves precision enhancement.
The transactions of the Institute of Electrical Engineers of Japan.C, 2007
The transactions of the Institute of Electrical Engineers of Japan.C, 2006
Multiagent learning to stabilize power grid is conducted. Each agent controlling each distributed... more Multiagent learning to stabilize power grid is conducted. Each agent controlling each distributed generator acquires suitable control action by using Q-learning. Simulation results show that the multiagent approach enables the grid to stabilize the voltage.
The transactions of the Institute of Electrical Engineers of Japan.C, 2006
Significances and approaches of applying intelligent systems to artificial electricity market is ... more Significances and approaches of applying intelligent systems to artificial electricity market is discussed. In recent years, with the moving into restructuring of electric system in Japan, the deregulation for the electric market is progressing. The most major change of the market is a founding of JEPX (Japan Electric Power eXchange.) which is expected to help lower power bills through effective use of surplus electricity. The electricity market designates exchange of electric power between electric power suppliers (supplier agents) themselves. In the market, the goal of each supplier agents is to maximize its revenue for the entire trading period, and shows complex behavior, which can model by a multiagent platform. Using the multiagent simulations which have been studied as "artificial market" helps to predict the spot prices, to plan investments, and to discuss the rules of market. Moreover, intelligent system approaches provide for constructing more reasonable policies of each agents. This article, first, makes a brief summary of the electricity market in Japan and the studies of artificial markets. Then, a survey of tipical studies of artificial electricity market is listed. Through these topics, the future vision is presented for the studies.
A New XGBoost Inference with Boundary Conditions in Real Estate Price Prediction
IEEJ Transactions on Electrical and Electronic Engineering
Real estate price prediction takes an important role in the economy that can drive up and down th... more Real estate price prediction takes an important role in the economy that can drive up and down the stock prices and even generate disruptive economic events. Many researchers have tried to understand the pricing mechanism with machine learning techniques such as support vector machine, neural network, random forest, and AdaBoost. The boundary problem, on the other hand, makes the pricing scheme more complicated, and this trend is accelerated especially in the situation of population decline in Japan. In this paper, we discuss how we could approach the boundary problem in real estate prediction. We propose a new comprehensive inference model extending and adapting XGBoost to the domain that has the boundary conditions problem by utilizing the distance between the instances in the domain data set to make the layers of bumpy boundaries smooth for more accurate predictions and robustness against the domain data set. The experiments result showed our proposed method performed well on bot...
Reinforcement Learning with High-Dimensional Parameters : Complex-Valued and Quantum Representation
Journal of the Society of Instrument and Control Engineers, 2012
Based Software Complexity Impact for Maintainability
AS-3-3 Information Support System Based on the Context Awareness Design by using the Active RFID Tags for Cognitive Disorders
A Study on Software Maintainability Evaluation using the Directed Graph Analysis
A Study on Japanese Prosodic Control Model In Restricted Speech
Journal of International Council on Electrical Engineering, 2017
We propose a visual-based self-localization system that requires less computational time and memo... more We propose a visual-based self-localization system that requires less computational time and memory resources when utilized in an indoor environment. In this study, localization is accomplished by comparing an observation with a significant amount of preprocessed images. The larger the database, the slower the self-location calculation. We constructed an economical database (codebook for local image features), in terms of computational time and memory usage, to resolve this challenge. We sequentially registered new codes to avoid duplication. Additionally, we indexed the codebook to speed up the calculation when using an approximating search. The experiment result illustrated that the strategies of reducing the size of the codebook and approximating the calculation contributed to reducing the calculation cost and improving the self-localization accuracy.
IEEE access, 2024
Diffusion models are emerging as a powerful solution for generating high-fidelity and diverse ima... more Diffusion models are emerging as a powerful solution for generating high-fidelity and diverse images, often surpassing GANs under many circumstances. However, their slow inference speeds hinder their potential for real-time applications. To address this, DiffusionGAN leveraged a conditional GAN to drastically reduce denoising steps and speed up inference. Its advancement, Wavelet Diffusion, further accelerated the process by converting data into wavelet space, thus enhancing efficiency. Nonetheless, these models still fall short of GANs in terms of speed and image quality. To bridge these gaps, this paper introduces the Latent Denoising Diffusion GAN, which employs pre-trained autoencoders to compress images into a compact latent space, significantly improving inference speed and image quality. Furthermore, we propose a Weighted Learning strategy to enhance diversity and image quality. Experimental results on the CIFAR-10, CelebA-HQ, and LSUN-Church datasets prove that our model achieves a state-of-the-art running speed among diffusion models. Compared to its predecessors, DiffusionGAN and Wavelet Diffusion, our model shows remarkable improvements on all evaluation metrics. Code and pre-trained checkpoints: .
A Routing Method using Link State Prediction for Wireless Smart-grid
Uploads
Papers by Tomoki Hamagami